A Study of Health Discrimination by Physical Examination for Labors: an Example of a Science Park in Taiwan

Abstract

There are three main issues in hospitals to be adddressed: personal health index estimation, health follow-up system establishment and health stability discriminated for examinees.
In this study, we took advantage of the Rasch model, using with the softwareWinsteps, to analyze the physical examination data from 11,830 labors in a Science Park in Taiwan. The measurement items were examined to see whether they fit the Rasch model and whether they measure a single construct in order to (1) estimate personal health index, (2) screen out unexpected results for further scrutiny, and (3) distinguish persons with Fit statistics to alarm unstable or bad health conditions.

The results showed that the all but one item fitted the model expectation reasonably well. Rasch analysis could be used to help physicians discriminate examinee health conditions by Fit statistics and personal health index estimation. We also found that the kidney, cancer disease and abnormal chest X-ray threaten females below 25 years old, symptoms of liver and gout diseases impacted the male, those 35 to 45 ages are suspect with coronary heart disease, 25 to 45 ages have the potential liver disease, and diabetes affected employees in factory of number two. The labors diagnosed with low health index by Rasch analysis were more likely to be threatened by liver diseases.

Keywords: Physical examination, Rasch Analysis, Single construct, DIF, Fit statistics.

以符合Rasch模式單一向度的試題內容，利用本研究挑出關鍵改善事項的5步驟（Opt in–DIF-Eye on-Evaluate-Recheck, ODEER）模式：(1)選入不滿意度大於0個logit的關鍵試題；(2)DIF偵測出影響關鍵試題之族群；(3)尋找關鍵試題變項上的其他潛在族群；(4)評估族群間滿意度與就醫動機強度間關係；(5)檢測知識結構映射圖中的就醫動機關聯效果，以彙整出改善措施與建議。